PROJECT SUMMARY/ABSTRACT Lung cancer, the number one cause of cancer-related deaths among both men and women in United States, can potentially be cured if diagnosed and surgically removed at Stage 1 of the disease. However, in 40% of individuals with Stage 1 non-small cell lung cancer (NSCLC, 90% of all lung cancer), surgery fails to prevent recurrence and cancer-related death within five years. Adding adjuvant chemotherapy on top of surgery for all Stage 1 NSCLC patients does not improve five-year survival in this population as a whole. Currently, we have no way to accurately predict which Stage 1 NSCLC patient is likely to die within five years. With this knowledge, we could perform targeted enrollment of only high-risk individuals into randomized clinical trials of adjuvant chemotherapy or other interventions in an effort to reduce cancer recurrence and improve five-year survival. The objective of this research is to prognosticate Stage 1 non-squamous NSCLC based on an analytically rigorous integration of multi-scale data (gene expression, tumor oncogene mutation, CT and PET imaging, and clinical) with the aid of didactic coursework and mentored training in data sciences. The central hypothesis is that an integrative signature incorporating multiscale data in Stage 1 non-squamous NSCLC will provide accurate prognostication of five-year survival. This project speaks to a long-term career goal of becoming an independent physician-scientist exploring the role of personalized medicine in neoplastic lung diseases. The first specific aim is to develop and evaluate an integrative genomic signature for Stage 1 non-squamous NSCLC prognostication by leveraging public domain data. A novel approach would be used, incorporating tumor oncogene mutation data (including KRAS mutation and EGFR amplification) alongside previously evaluated clinical and gene expression data in order to develop a more comprehensive prognostic signature for five-year survival. The second specific aim is to develop and evaluate an integrative radio-genomic signature for Stage 1 non-squamous NSCLC prognostication by leveraging a local patient cohort that is heavily annotated with radiographic and genomic data. Capitalizing on the sponsoring faculty's expertise in studying the associations of molecular data with imaging features, the novel approach here would be to develop and evaluate a new prognostic signature incorporating specific volumetric CT features and PET markers of tumor metabolic activity into the most robust genomic signature currently available. The contribution of the proposed research is expected to be the provision of an accurate, multiscale prognostic signature that would improve clinical decision-making. This contribution will be significant because it is expected to have broad implications for the utilization of novel personalized medicine approaches towards accurate prognostication and treatment of NSCLC, with methodological application to all types of cancer.